crossplane

Pass

Audited by Gen Agent Trust Hub on Jun 29, 2026

Risk Level: SAFEPROMPT_INJECTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
  • [PROMPT_INJECTION]: The skill has an indirect prompt injection surface as it is designed to process user-defined input parameters in Crossplane Claims to orchestrate cloud infrastructure.
  • Ingestion points: User-provided values in YAML Claim manifests (e.g., SKILL.md, references/enterprise-patterns.md).
  • Boundary markers: The skill encourages the use of XRD OpenAPI schemas to define strictly typed parameters and constraints.
  • Capability inventory: High-privilege cloud resource management (RDS, VPC, EKS) via Crossplane providers.
  • Sanitization: Uses OpenAPI validation (minimum, maximum, enum, pattern) and custom logic within Go/Python/TypeScript composition functions to validate inputs.
  • [EXTERNAL_DOWNLOADS]: The skill references and installs Crossplane providers and configuration packages from the Upbound OCI registry and other OCI-compliant registries.
  • Evidence: References to xpkg.upbound.io/crossplane-contrib/provider-aws:v1.1.0 in SKILL.md and various providers in references/package-management.md.
  • [COMMAND_EXECUTION]: Includes a utility Bash script to monitor the health of Crossplane providers using standard CLI tools.
  • Evidence: SKILL.md contains a script using kubectl and jq to query the status of installed providers.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 29, 2026, 08:46 PM
Security Audit — agent-trust-hub — crossplane